具有非线性输入环节系统的自适应递归输出反馈控制研究

Nonlinear input problems such as hysteresis, dead-zone and friction exist widely in industrial systems. Also most practical systems have model uncertainty and the unmeasured states. Therefore, the adaptive output feedback control for the nonlinear system with nonlinear input has drawn considerable attention. This research project aims to explore some important open problems of tracking control for nonlinear systems with nonlinear input by utilizing the adaptive recursive output feedback strategy. Specially, this proposal focuses on the following: 1) The design of global adaptive output feedback control for nonlinear system involving input nonlinearity, unknown parameters and the nonlinearities dependent on the unmeasured states and its application in robot tracking control system; 2) The design of global adaptive output feedback control for actuator failure system with nonlinearity input and its application in robot tracking control system. In addition, the contol scheme can guarantee the steady state and transient state tracking performance of actuator failure system without increasing control complexity; (3) The design of robust adaptive output feedback control for nonlinear time delay system with nonlinear input and unknown signs of the high frequency gains and its application in the metal cutting system. The Nussbaum method is abandoned in this control scheme and the signs of the high frequency gains are directly estimated for the purpose of removing the rigorous assumption in the existing literatures. The results of the project will provide basic theory and methods for the control of industrial systems with nonlinear input. And it has important scientific and engineering significance.

工业系统常见磁滞，死区，摩擦等非线性输入问题，并且实际系统往往具有模型不确定及大部分状态不可测性，故具有非线性输入系统的自适应输出反馈控制受到了极大关注。本项目针对具有非线性输入环节非线性系统跟踪控制中的关键问题，拟开展自适应递归输出反馈控制研究：1）同时含有非线性输入环节、未知参数、与不可测状态相关非线性项系统的全局自适应输出反馈控制设计并应用于机器人轨迹跟踪系统中；2）考虑非线性输入环节执行器故障系统的全局自适应输出反馈控制设计，在不增加控制复杂度的情况下，保证执行器故障系统的稳态及瞬态跟踪性能并应用于机器人轨迹跟踪系统中；3）考虑非线性输入的高频增益符号未知时滞系统的鲁棒自适应输出反馈控制设计，放弃Nussbaum方法，直接估计高频增益符号，消除已有方案中的一些苛刻条件并应用于金属切割系统中。本项目研究成果将为含有非线性输入工业系统的控制提供理论基础和方法，具有重要的科学及工程意义。

非线性输入问题广泛存在于工业系统中，典型的如磁滞、摩擦、死区及间隙等，其严重影响了系统的控制性能及稳定性，并且实际物理系统往往具有模型不确定及参数未知性。本课题以具有非线性输入环节的几类典型非线性系统为研究对象，以机器人轨迹跟踪控制系统为相应的实验验证平台，解决鲁棒自适应递归控制中的如下关键问题：1）同时含有非线性输入环节、未知参数、与不可测状态相关非线性项系统的全局自适应输出反馈控制设计；2）考虑非线性输入纯反馈系统的鲁棒递归控制设计，在不增加控制复杂度的情况下，保证纯反馈非线性系统的精确跟踪性能，其适用于部分执行器故障系统中；3）考虑非线性输入的高频增益符号未知系统的鲁棒自适应控制设计，放弃Nussbaum方法，直接估计高频增益符号，消除已有方案中的一些苛刻条件，保证了闭环系统的稳态及瞬态跟踪性能。